A young scientist with curly, reddish hair tucked beneath a knit cap stepped gingerly onto the three-day-old ice of a remote lake in northeastern Siberia. Coating the black depths like cellophane, the thin film held no promise to bear her weight, but a sudden dunk in the frigid water was a risk she had to take. Searching the lake by rickety rowboat all summer had failed, and any day winter’s first big snow would engulf the region, obscuring the lake’s surface until spring. She could not afford to wait that long.

The woman shivered in her worn, blue down jacket and glanced up at the overcast sky. After one more cautious step, she spotted her quarry: a cluster of platter-size bubbles frozen into the ice. Those pockets of gas, which had risen from thawing permafrost—formerly frozen soil—at the lake’s bottom, were the aim of her doctoral research. Long elusive, they suddenly stood out like white stars against a night sky, though less serenely. With a small pick she cracked the icy skin of one of the bubbles and remained unfazed when it hissed back like a punctured gas pipe. Leaning forward, she apprehensively struck a match just above the broken bubble and flames as high as her head burst skyward. The flammable substance was methane, a greenhouse gas that could cause more global warming than carbon dioxide (CO2).

Today, nearly seven years after igniting that first bubble, Katey Walter finds herself center stage in an environmental drama playing out across the frozen north. Now a 33-year-old assistant professor at her alma mater, the University of Alaska–Fairbanks, Walter was the first to explain the mysterious methane emissions from Arctic lakes. She isn’t shy about touting their significance as a ticking time bomb. In a complete Arctic thaw, these lakes could discharge a whopping 50 billion tons of methane: 10 times the amount already helping to heat the planet.

Whether a total or more moderate release is in store is still anyone’s guess. But pound for pound, methane in the atmosphere traps 25 times more of the sun’s heat than CO2 does. Consequently, even a modest thaw of the perennially frozen soil that lies under these ephemeral lakes and caps the dry land around them could trigger a vicious cycle: warming releases methane and creates lakes, which thaw permafrost and liberate more gas, which intensifies warming, which creates more lakes, and so on. Some Arctic lakes are growing larger, and researchers are eyeing them suspiciously as a reason why global methane concentrations shot up in 2007 and have stayed high ever since. Other signs indicate that permafrost thawing on the Arctic seafloor may be loosening the cap on large pockets of methane stored deeper down.

Walter is sounding the alarm even louder than before because global warming is taking a special toll across the far north. The region is heating up twice as quickly as the rest of the globe, rapidly melting sea ice in the Arctic Ocean as well as the permafrost, which underlies 8.8 million square miles of the Northern Hemisphere. Leading climate models already suggest greenhouse warming as a result of most of the Arctic’s permafrost thawing by 2100—and the estimates do not yet include the potentially vast additional warming imparted by methane bubbling up out of chilly waters. Walter and others are trying to determine just how much methane could be released into the atmosphere, how soon, how aggressively that release would accelerate the earth’s warming and whether anything can be done to temper the escalating threat.

Burps and Belches
Scientists know with great certainty how much methane is in the earth’s atmosphere at any given time from sampling its concentration weekly at dozens of sites worldwide. By plugging these measurements into global climate models, they know methane is responsible for a third of the current warming trend. Exactly how much gas comes from where is harder to say, which is why the Arctic lake bubbles were so long overlooked.

Methane is emitted anywhere organic matter ferments—be that a cow’s belly or frozen soil that starts to thaw. Permafrost, which averages 80 feet thick, is chock-full of dead plant and animal matter that has been locked in cold storage for thousands of years. Conventional wisdom long held that permafrost should take thousands of years to melt away, so researchers expected it to play a negligible role in climate change. But recent findings—Walter’s lake discovery in particular—have wrecked that prediction.

Walter’s work revealed that the relatively warm lake bed was indeed thawing the frozen earth directly below it, down several dozen feet. Thawing a block of permafrost is like taking a package of frozen hamburger out of the freezer and leaving it on the kitchen counter. As the meat warms, ravenous microbes consume it, giving off a gas as a by-product. On dry land, microbes convert the dead animal and plant matter primarily into CO2. But in the wet, oxygen-starved depths of a lake, they instead release methane. Walter’s best guess is that researchers have been underestimating methane emissions from Arctic wetlands by as much as 63 percent.

This methane alert, which Walter raised first in her doctoral thesis, captured the attention of the U.S. Council of Graduate Schools, which in 2006 granted her the nation’s most prestigious honor for doctoral dissertations. She credits her discovery to living lakeside from one season to the next. Most scientists tend to be in the field only during the summer, when the bubbling seeps of gas are hard to spot in open water, or in the winter, when the lake is buried under six feet of snow. The same camouflage deterred Walter until that overcast afternoon in October 2002, when she decided to remain lakeside during Siberia’s brief transition from summer to winter. By the spring of 2003 she knew exactly what she needed to do: place her gas traps directly over known seeps. Her results have since riveted attention on how drastically thawing permafrost could speed up global warming.

Trapping the Demon
During four years of doctoral work, Walter spent 20 months in the Siberian wilderness, often alone or with only one loyal field assistant. She hiked up to eight miles a day across sodden tundra and braved icy waters on several occasions, deliberately as well as accidentally. She knew exactly what she was getting into; as a high school exchange student to Russia nine years earlier, Walter had learned the language and was deeply touched by the harsh conditions of post-Soviet life. She jumped at the chance to return.

Headquarters for much of those four years was the so-called Northeast Science Station, a small outpost in Cherskii, about 90 miles south of the Arctic Ocean. The station’s director, legendary ecologist Sergey Zimov, who had published his suspicions about the role of lake emissions in climate change, helped to define Walter’s straightforward goal: find a way to quantify the methane release and determine what fraction could be attributed to thawing permafrost. Walter’s first challenge was to invent a way to capture the gas. She knew she would eventually need hundreds of contraptions to adequately sample the two large lakes in her study, so her design needed to be simple—and cheap. Walter and her Siberian field assistant spent weeks cobbling together traps in the station’s cramped attic, mostly from recycled trash they found at abandoned Soviet military bases and along dusty dirt roads. For each trap, they secured an inverted plastic bottle to the center of an umbrella-shaped plastic skirt, which was held open by a hoop of wire to funnel bubbles upward. They made 75 of them in all.

During her first excursions, Walter dutifully placed the traps randomly across the lakes’ unfrozen surface, according to standard scientific protocol. “We put a lot of hard work in that, and I was frustrated,” Walter says. “We could see the bubbles, but we weren’t catching much gas.” It wasn’t until she walked out onto the freshly frozen ice for the first time and saw the disparate collections of bubbles that she realized the methane was rising up at discrete points. She made an executive decision, a bit nervously and without the consent of her thesis advisers, to set the traps directly over seeps when the lakes thawed the following spring. So, in 2003, she set about anchoring many of her traps right near the lake bottom—a job that called for a snorkel and wet suit. Locating a seep and setting the necessary tripod of weights and ropes for a single trap required two and a half hours submerged in lake water still gripped by winter chill.

That same spring Walter’s colleague at Fairbanks, Vladimir Romanovsky, whose computer simulations are some of those predicting a dramatic thaw this century, made an unrelated visit to Cherskii and observed Walter swimming in the icy lakes: “She’s a tough girl,” he says simply. By summer Walter found herself in the hospital with pneumonia. “But a couple of months later, and with a good dose of Russian antibiotics, I was back in the lakes,” she recalls. When winter came again, the drill changed from snorkeling to shoveling. For hours at a time Walter dug away at the snow atop the ice, clearing paths above the seeps and marking them with flags as she went. “The Siberians were laughing their heads off at how much money we were spending to come to Siberia to shovel snow off the frozen lakes,” she says. But no one was laughing when the world learned about her hard-won findings.

Proof for a Spike
Walter’s intimate relationship with a handful of Siberian lakes initially brought lake emissions into the limelight, but it was her analysis of their global importance, which she reported in two major scientific journals in 2007, that really turned heads. The potential for emissions to increase dramatically became clear through her work with paleoecologist Mary Edwards of the University of Southampton in England, who has studied the life histories of Arctic lakes. Together they showed that methane bubbling out of Arctic lakes could have been responsible for up to 87 percent of the spike in methane emissions that helped the planet warm from the most recent ice age. At that time, roughly 11,400 years ago, global methane concentrations rose 50 percent in less than 200 years.

Many scientists are keen to determine whether such a dramatic spike might happen again. A steady march of global warming, spread out over hundreds or thousands of years, could set off the gaseous Arctic time bomb slowly. But a quicker thaw could ignite a runaway outgassing of methane.

For about a decade it has been clear that the ongoing loss of sea ice is accelerating the Arctic’s rapid warming, says climate modeler David Lawrence of the National Center for Atmospheric Research in Boulder, Colo. When summer ice retreated to a record minimum in 2007 and again last year, the outlook seemed to worsen by the month. New estimates, published in April by the National Oceanic and Atmospheric Administration, predict nearly ice-free summers by 2037—three times sooner than earlier models indicated. The prospect of more open water has nations scrambling to stake oil and gas claims to the Arctic seabed [see “Arctic Landgrab,” by Jessa Gamble; Scientific American Earth 3.0, Vol. 19, No. 1, 2009], but the backlash for climate change could be severe. Dark seawater absorbs more of the sun’s heat than white ice does, thus warming the region’s air and thereby the soil, putting permafrost at risk. Lawrence’s newest global climate simulations predict that warming associated with spells of particularly rapid loss of sea ice could lead directly to faster permafrost thaw. During such episodes, which would last five or 10 years, autumn temperatures might increase by as much as nine degrees Fahrenheit along Arctic shorelines, and the heat penetrating inland would more than triple the average warming rates previously assumed.

Rising inland temperatures are fueling another dramatic change “potentially as profound as the loss of sea ice,” says Matthew Sturm of the U.S. Army’s Cold Regions Research and Engineering Laboratory in Fort Wainwright, Alaska. Shrubs are taking over great swaths of the tundra. During the summer, shrubs absorb more sunlight than does the mossier, grassier vegetation they replace, warming the ground further. And in the winter they create snowdrifts that help the ground hold on to summer heat. Sturm’s extensive comparison of 6,000 aerial photographs taken across northern Alaska for oil exploration during the 1940s to present-day surveys of the same locations shows significant shrub expansion, now covering 77,000 square miles.

Double the Emissions
Some Arctic scientists are quick to point out that certain environmental changes could slow warming rather than speed it up, however. Sturm has also found areas where shrubs are not expanding and soils are colder. In other regions, the conversion of mossy tundra to thin forest, not shrubland, offsets some rise in greenhouse gases by storing more carbon in trees. Still, the consensus is that warming will dominate. “The question is whether this is a weak positive or a strong positive,” Lawrence says. “It may take a long time to get the numbers right.”

Even now, though, Lawrence is willing to offer a lower bound of methane release. It is easy, he says, to envision conditions that double methane emissions from the Arctic by the end of the 21st century simply by activating more microbes in those uppermost few feet of Arctic soil that thaw every summer, the so-called active layer. But more lakes formed because of thawing would send that estimate skyrocketing. Walter’s work suggested that the lakes near Cherskii expanded significantly between 1974 and 2000 and that as they did, they ate into the permafrost along their shorelines. She found that methane rises up most vigorously at these outer edges, which fueled her 2006 estimate that an expansion of thaw lakes increased methane emissions in the region by 58 percent.

Going back to Siberia as a professor in 2008 made her wonder if it was time to update that estimate. The banks of the lake where she lit her first methane bubble in 2002 had advanced greatly. “The dramatic changes to my study sites really made my eyebrows go up,” she says. “I couldn’t even recognize the lake margins. Some ponds appeared to have doubled or tripled in size.” If other lakes experienced similar growth they may have contributed to the global methane spike that began two years ago.

Walter still spends four months or more each year walking the lakes in dogged pursuit of answers. Her collaborators and study sites have expanded considerably. To date, she and her colleagues have visited 60 lakes in Siberia and Alaska, but that is still only the tip of the proverbial iceberg. With no hope of visiting every Arctic thaw lake in person, her team is now working on a technique to spot methane seeps from space. One new high-resolution German satellite, TerraSAR-X, is making it possible to identify distinct patches of bubbles on the surfaces of frozen lakes—and to keep track of which patches are growing.

One More Reason to Cut Back
Once a given helping of permafrost starts to thaw and a gas leak starts, not much can be done about it. Local villages could capture the bubbling methane gas and use it to replace diesel fuel (technologies already exist for capturing methane released from landfills). But that is a “very small fix,” Walter concedes. The only real solution is to slow the thaw itself.

Those of us living at lower latitudes can make the greatest difference. The model linking permafrost thaw to loss of sea ice predicts that both processes could be slowed considerably if humanity stabilizes CO2 emissions soon, slowing the atmospheric warming that is generating the methane. “It’s not a runaway train,” Lawrence says. Not yet, anyway, Walter and others warn. Lawrence sounds optimistic when he says, “Perhaps we have to reduce emissions by 80 percent rather than 70 percent by 2050.” But such dramatic reductions won’t be easy. Since 2000 human activities have raised CO2 concentrations much faster than expected.

Even if humanity finds the resolve to slow warming, too much thawing in the wrong place could tap submerged caverns of methane. Just below the permafrost layer in many locations lurk large pockets of pure gas that formed millions of years ago. Some of the pockets are run-of-the-mill natural gas reserves, but others are so-called methane hydrates, massive deposits of ice that contain large amounts of gas within their crystalline structure.

Some scientists suspect that permafrost acts as a cap that protects hydrates from melting, particularly in the shallow Arctic seafloor, where the hydrates are found only a few tens of feet deep. The more that sea or lake waters thaw the permafrost below, the more likely this cap is to blow suddenly, releasing jets of methane up through the water and into the atmosphere. A team that included two of Walter’s colleagues at Fairbanks found such plumes rising up from the shallow continental shelf of Siberia in 2008. Possibly, these releases have been happening for a long time, and we are only now noticing them. But the discoverers point out that the Siberian shelf alone holds an estimated 1.4 trillion tons of methane in the form of gas hydrates—equivalent to the newest estimates of the total greenhouse gases that would be released during a complete permafrost thaw.

Many researchers note that methane hydrates exist below the permafrost on land as well. The deposits are generally assumed to be too deep to be at risk of thawing. But that assumption, like others before it, has been cast in doubt. If Walter confirms indications from a field excursion earlier this year that Arctic lakes are tapping a methane source even older and greater than permafrost, her alerts would have to be cranked up considerably. And those bubbles she lights would take on an even more sinister glow.

More beasts, less burden: Large animals could help keep permafrost frozen.
Strangely enough, one way to slow the thawing of permafrost is to reintroduce massive herds of large, plant-eating animals to the Arctic landscape to mimic the days when millions of mammoths roamed the Siberian steppes. Although the idea may sound like science fiction, it is based in sound ecology. “Snow is like a down jacket that keeps the ground warm,” University of Alaska–Fairbanks researcher Katey Walter points out. “As the activity of animals compresses the snow or removes it through their foraging, the cold winter temperatures can penetrate deeper into the ground and keep the permafrost frozen.” Indeed, ecologist Sergey Zimov, director of the Northeast Science Station in Cherskii, Siberia, has hired local villagers to fashion that solution with their own hands. They have fenced off a 625-square-mile ranch Zimov calls Pleistocene Park and stocked it with moose, reindeer and Yakutian horses. Zimov has mimicked mammoths by driving around a military tank to crush the ground, too. He argues that the climate is still optimal for grassland, which would also insulate the permafrost below, if animals can thrive to cultivate it. Hunting, not climate, he points out, is blamed for the mammoth’s demise.

Note: This article was originally printed with the title, "The Peril below the Ice."